Qwen3.8 2.4T A95B
alibaba/qwen3.8-2.4t-a95b- Alibaba
- Family · Qwen3
- Released 12 Aug 2026
- Open weights
- Text → Text
- Reasoning · adjustable
- Tools · JSON
Not on OSMEXO yet
Spec & list price
Context1M
Max output128K
Input / 1M$2.00
Output / 1M$6.003× input
Scores · standing
Intelligence39.9#34/489
Coding71.9#20/196
Usage
Tokens / day4.2B▲ 17% 7d
Usage rank#125
Requests / day212.2K
Tool calls / req0.49
Per request · 30 days
Avg input44.3K3.0× median
Avg output2.0K2.9× median
Cache hit87%1.3× median
Reasoning share80%1.4× median
Tool-call failures1.9%about median
Consumption
Peak 24.1B 14 OctLow 1.7B 17 OctRank today #125
News & trends
195 mentions this week ▲ 70%Qwen3 family · videos, posts, articles · 30 days
Trending · 14 days · Qwen3 family
- Memdoor: A team of 5 Qwen3.8-27B models beats Fable 5 on LiveCodeBench Hard370 points
- Neuron Explorer for Qwen3.5-4B: See every neuron at work, make targeted edits26 points
- New Card for Qwen3.8 27b and MiniMax H3567 upvotes
- Qwen 3.8 27B vs Strata Coder - Local LLM battle75.2K views
- Qwen 3.8 27b is still enough for almost anything445 upvotes
- DavidAU Qwen3.8 27B Turbo Brilliance 50X-Reasoning... tested - Local LLM setup87.7K views
News · latest from Qwen3
- Memdoor: A team of 5 Qwen3.8-27B models beats Fable 5 on LiveCodeBench Hardgithub.com · 31 Oct
- Neuron Explorer for Qwen3.5-4B: See every neuron at work, make targeted editsgithub.com · 26 Oct
- How to train your own Jev for $17Together AI · 23 Sep
Benchmarks
bar = share of the top score · Benchmark data: Artificial AnalysisSpeed · median across 5 providers, latest day
Output speed#292/33330 tok/s2.2× slower than median
Fastest provider83 tok/sbest of 5 providers
First token#2381.75 s1.7× longer than median
First answer token#9254 sthinking included
#12 cheapest of 19 with a similar score
Value
Intelligence Index
Blended price per 1M tokens (log)
Family progress
Intelligence Index
- Family releases33.7
Technical details
- Reasoning
- on · xhigh
- Tokenizer
- Qwen
- License
- other
- Parameters
- 2.4T
- Moderation
- none
- Downloads · likes
- 25.9K · 1.4K
Reasoning effort
- xhigh · default
- medium
- low
Request parameters 20
- tools
- tool_choice
- response_format
- structured_outputs
- reasoning
- reasoning_effort
- include_reasoning
- temperature
- top_p
- top_k
+10 more
- min_p
- max_tokens
- stop
- seed
- frequency_penalty
- logit_bias
- logprobs
- presence_penalty
- repetition_penalty
- top_logprobs
When to use
Best for
- Hard science and expert-level reasoning
- Coding agents in terminal environments
- Long document analysis over very large context
- Banking and customer-service tool workflows
Think twice for
- Image or video input (text-only)
- Low-cost bulk classification
- Latency-sensitive simple chat
- Running on small self-hosted hardware
Same family
Qwen3- 6.0
- 7.2
- 6.9
- 6.1
- 5.2
- 13.1
Alternatives
other vendors · similar capability- S43.7
- 40.9
- 36.0
- 38.2
Want Qwen3.8 2.4T A95B on OSMEXO?
Sign in to request it. Requests are reviewed; a request does not guarantee availability.